llms.txt Content
# Sandbox0
Always available. Without always-on compute costs.
Sandbox0 provides managed AI agent sandboxes so you can give every user their own cloud computer. Preserve the entire writable root filesystem across successful checkpoints, including system configuration, installed software, dependencies, and user files. Growth-enabled templates expand RootFS automatically online within the configured device ceiling while the agent keeps running; existing images and snapshots retain their captured geometry. Pause compute between tasks and resume on demand. On Sandbox0 Cloud, paused sandboxes incur no compute charges; idle sandboxes that remain running still accrue charges. Default pause/resume preserves the filesystem and starts fresh processes. Experimental memory checkpoints can also restore supported process state. With auto_resume enabled, supported inbound access can wake a paused sandbox; resume takes time and requires available capacity. Incremental checkpoints and demand reads avoid full-RootFS copies. Sandbox0 also provides network control, snapshots, and forks.
Sandbox0 Cloud is the supported service. Self-hosting is no longer supported. Bring Your Own Cloud (BYOC) is planned and is not available yet.
Deployment options: https://sandbox0.ai/docs/sandbox/deployment.md
## s0 CLI
When a user asks you to set up Sandbox0, install the `s0` CLI if it is not already available, use it for Sandbox0 operations, and consult the agent-readable docs below as needed.
Install on Linux or macOS:
```bash
curl -fsSL https://raw.githubusercontent.com/sandbox0-ai/s0/main/scripts/install.sh | bash
```
Install on Windows PowerShell:
```powershell
irm https://raw.githubusercontent.com/sandbox0-ai/s0/main/scripts/install.ps1 | iex
```
Install with Go:
```bash
go install github.com/sandbox0-ai/s0/cmd/s0@latest
```
Authenticate with `s0 auth login` for interactive CLI use. The CLI starts device login; ask the user to approve it when required. If no team is selected after logi